Filtered Views from Voxel51 - October 2023
Welcome to the Filtered Views, the monthly newsletter for customers, community members, and fellow AI enthusiasts who are looking to stay up-to-date with recent happenings in computer vision and FiftyOne. Need to reach us? Drop us a line in the FiftyOne Community Slack.
🏆 Success stories
Ever wondered how others are using FiftyOne and what benefits they see? Here is this month's highlight!
“At Protex AI we develop a platform to monitor worker health and safety using existing CCTV infrastructure. Our core computer vision technologies are object detection, classification, and pose estimation. We use FiftyOne as a vital component in our pipeline to validate dataset annotations, intelligently subsample datasets to ensure balance, and also to visualize and debug model predictions to assess accuracy.” -- Patrick Rowsome, Lead Computer Vision Engineer, Protex AI
See more stories from people and organizations building remarkable machine learning and AI using FiftyOne and FiftyOne Teams.
Is your organization using FiftyOne to solve interesting computer vision problems? Share your story and claim a box of goodies as a thank you!
💻 New products
Here are the latest point releases to the open source and Teams versions of FiftyOne, along with links to learn more:
Open source FiftyOne
FiftyOne Teams
🔥 New features
September’s releases brought us three big exciting features!
Dataset versioning
Dataset versioning in FiftyOne Teams allows you to track the lineage of your data, view the exact version of a dataset on which a given model was trained, protect against accidental data modification, and much more. With dataset versioning, you can trust FiftyOne Teams as the single source of truth for your organization’s data. See it in action in this one minute video!
Native Ultralytics integration
FiftyOne now includes a native Ultralytics integration, making it easy to train the latest Ultralytics models like YOLOv8 on your FiftyOne datasets with just a few simple lines of code.
Delegated operations
FiftyOne Teams adds a powerful new Delegated Operations feature to FiftyOne’s Plugin framework that allows you to schedule builtin and/or custom tasks from within the App that are executed on a connected workflow orchestrator like Apache Airflow.
Why is this awesome? Your AI stack needs a flexible data-centric component that enables you to organize and compute on your data. With Delegated Operations, FiftyOne becomes both a dataset management/visualization tool and a workflow automation tool that defines how your data-centric workflows like ingestion, curation, and evaluation are performed. In short, think of FiftyOne Teams as the single source of truth on which you co-develop your data and models together.
📣 Upcoming events
Oct 16-17: PyTorch Conference 2023
Oct 19: Computer Vision Summit Boston
Oct 25-26: MLOps World 2023 Austin
...
📙 New resources: blogs
Introducing Badger! An open source Python library that allows you to create, edit, store, and manage custom GitHub badges, all from the command line.
New FiftyOne Plugin to pre-label your computer vision data with CLIP, SAM, and other zero-shot models!
Recommended by LinkedIn
New tutorial: Combine FiftyOne and Ultralytics YOLOv5 to get a robust solution for training drone data.
Two new FiftyOne Plugins! These two plugins effectively allow you to query documents like pages of old books, handwritten notes, or resumes by the text that they contain.
Use FiftyOne to explore nuScenes, a public large-scale dataset for autonomous driving.
Explore a new benchmark dataset for calving front delineation in SAR images - CaFFe, and try it in FiftyOne.
New FiftyOne Plugin to deduplicate your entire dataset from within the FiftyOne App, without writing a single line of code.
Everything you need to know to get started working with the new FACET dataset by Meta so the models you deploy are not learning biases.
...
🎥 New resources: videos
...
👋 Stay connected
Open source software like this doesn’t happen without an awesome open source community backing it. If you haven't yet, we'd love to have you as part of the community!